26. Verbal Aspect1
Henriëtte de Swart
1. The Role of Verbs and Arguments in the Grammar of Aspect
Semantic studies are often hard to carry out, for the distinctions are subtle, and intuitions not
always easy to grasp. In the study of aspect, this problem is compounded by the complexity
and abstract nature of the theoretical concepts, which often have a long history in languagespecific grammars and in the linguistic literature (see Binnick, 1991, and Filip, to appear, for
historical overviews). This first section offers some observations about English, which will set
the scene for the exploration of aspect in a cross-linguistic perspective in the remainder of this
chapter.
Verbal tense, which Comrie (1985, p. 1) defines as the “grammaticalization of location
in time,” commonly serves in natural language to anchor the situation described by the
sentence to the time axis (cf. Hewson, this volume). Thus the distinction between (1a), (b) and
(c) is temporal in nature:
(1)
a.
Bill was in love with Susan.
b. Bill is in love with Susan.
c.
Bill will be in love with Susan.
Tense is deictic and requires reference to the speech situation. The tenses in (1) locate the
situation of Bill’s love respectively before, at (or around), and after the speech time. In
contrast, the distinction between (2a) and (b) is aspectual in nature:
(2)
a.
Sarah wrote a dissertation in 2009.
It was completed in September.
1
#I think she is still working on it.
#She never finished it, for she died in September of that year.
b. Sara was writing a dissertation in 2009.
It was completed in September.
I think she is still working on it.
She never finished it, for she died in September of that year.
Both sentences relate to the past, but in (2a) the writing of the dissertation is presented as a
completed event: no writing is going on at the speech time anymore, and the dissertation is
finished. (2a) cannot be followed by sentences that deny completion of the event in the past.
(2b) presents the writing process as ongoing: the dissertation is under way at some point in the
past, but is not finished yet. The possible follow-up of (2b) allows completion of the
dissertation before now, or a continuation of the writing into the present, or a termination of
the process without completion. The difference between (2a) and (2b) is due to the presence
of the Progressive form in (2b), and its absence in (2a). According to Comrie (1976, p. 3)
“aspects are different ways of viewing the internal temporal constituency of a situation.”
Grammatical aspect is therefore also called “viewpoint aspect” (Smith, 1991/1997). Aspect is
not inherently deictic, and it does not anchor the situation to the time axis. Aspect may
however affect temporal structure, as is clear from the sentences in (3):
(3)
a.
When Bill came into the office, Sara left through the back door.
b. When Bill came into the office, Sara was leaving through the back door.
We normally perceive (3a) as reporting two events that follow each other in time: Sara left
just after Bill came in, perhaps as a reaction to his arrival. (3b) describes the two situations as
2
overlapping in time, perhaps as an explanation of why Bill missed seeing Sara at the office.
The close connections between aspectual and temporal structure motivate the study of tense
and aspect in conjunction.
The progressive construction in (2b) and (3b) is a grammatical aspect marker, as it is
part of the verbal inflection system of English. Grammatical aspect is distinct from lexical
aspect, also called Aktionsart, actionality, aspectual class or situation aspect (see Filip, this
volume). Lexical aspect bears on inherent features of the verb, as we see in (4). All three
sentences are in the simple past, and they contain no overt aspectual markers. Yet, they
describe situations with rather different internal temporal constituencies:
(4)
a.
Bill was in love with Susan.
b. Sarah wrote a dissertation.
c.
Carl reached the top of the mountain.
Sentence (4a) (=1a) describes Bill as being in a certain state of mind. This state
protracts over time, but nothing is occurring. (4b) (=2a) describes a completed event. Writing
a dissertation is a process that requires a certain amount of time, but it has an inherent
endpoint: the event is completed when the dissertation is finished. Reaching the top requires a
long preparatory stage of working one’s way up, but the event of reaching the top itself is a
momentaneous transition from the stage of working one’s way up to the resultant state of
being at the top. The aspectual differences between (4a-c) are due to the choice of the verb.
This is what opposes lexical aspect to grammatical aspect.
Verkuyl (1972) labels the term lexical aspect a misnomer, as the verb is not solely
responsible for the aspectual character of the sentence. The pair of sentences in (5) illustrates
this:
3
(5)
a.
Susan ate an apple.
b. Susan ate apples.
Sentence (5a) describes a completed event (the apple is finished), whereas (5b)
describes an unbounded process. Some apples must have been consumed to make (5b) true,
but the unspecified number of apples does not define an inherent endpoint. Verkuyl argues
that the semantics of the noun phrase contributes to the aspectual character of the sentence as
a whole, and aspect needs to be defined at the level of the predicate-argument structure (VP
and S). This insight leads him to prefer the term aspectual class or situation aspect. Verkuyl’s
intuition that subjects and objects play a role in creating temporal structure and imposing
boundaries on the situation is worked out in various analyses (Dowty, 1979; Krifka, 1989,
1992; Verkuyl, 1993; Jackendoff, 1996).
Garey (1957) characterizes examples like (5a) as telic, and examples like (5b) as
atelic. Telic and atelic verb phrases give rise to different inference patterns, as illustrated in
(6a, b):
(6)
a.
Susan was drinking wine → Susan drank wine.
b. Susan was drinking a glass of wine →
/ Susan drank a glass of wine.
English has a small set of verbs that are inherently telic and necessarily require a
delimitating argument, such as eat up, drink up. They are therefore incompatible with a bare
plural or bare mass noun:
(7)
a.
Susan ate up *cake/the cake/*apples/the apples.
4
b. Susan drank up *wine/the wine/*glasses of wine/two glasses of wine.
The contrast between (5) and (7) shows that the thematic relation between the verb
and its arguments is relevant to the aspectual characterization of the sentence (Krifka, 1989,
1992).
Aspectual class and grammatical aspect are independent theoretical notions, but there
are clear interactions between them. For instance, the English Progressive does not easily
apply to stative verbs (8a), or creates special meaning effects when it does, as in McDonalds’
slogan (8b):
(8)
a.
??Bill was being in love with Susan.
b. I’m lovin’ it!
In section 2, we briefly discuss the main terminological distinctions that are drawn in
the literature on aspectual class (section 2). There is more to grammatical aspect than the
English Progressive, so the debate on aspect has to be situated in a broader cross-linguistic
perspective (section 3). Sections 4 and 5 are concerned with the compositional interpretation
of aspect.
2. Aspectual Classifications: Basic Distinctions and Challenges
For non-stative verbs, argument structure drives the telic/atelic distinction (cf. example 5).
Krifka (1989, 1992) analyzes for-adverbials as expressions that measure the duration of a
situation without clearly established beginning- and endpoints (9a). In-adverbials measure the
time it takes to complete an event with an inherent endpoint, as illustrated in (9b). Foradverbials felicitously combine with atelic verb phrases, but disprefer telic verb phrases (9c).
5
Conversely, in-adverbials combine with telic verb phrases, but disprefer atelic verb phrases
(9d):
(9)
a.
Susan wrote letters/drank wine for half an hour.
b. Bill wrote a dissertation in six months/ drank a glass of wine in five minutes.
c.
??Bill wrote a letter for an hour/ drank a glass of wine for an hour.
d. ??Susan wrote letters/drank wine in half an hour.
A process of aspectual coercion (see section 4.4 below) often renders it possible to
make sense of such dispreferred combinations. For instance, (9c) can be read as ‘working on
writing a letter for an hour’, or ‘drinking from a glass of wine for an hour’.
The for/in-criterion shows that a broader class of verb phrases can be classified as telic
or atelic than those in (5) and (7). Stative verbs like be in love and activities like waltz qualify
as atelic according to (10a) and (b), whereas verb phrases describing instantaneous events like
reach the top are telic (10c):
(10)
a.
Bill was in love with Susan for/*in many years.
b. At the wedding, Bill waltzed for/*in several hours.
c.
Bill reached the top of the mountain in/*for two days.
The common feature of states (like be in love, 10a) and activities (like eat apples, but
also swim, waltz, push a cart, 10b) is that they describe unbounded situations without an
inherent endpoint. In terms of their interpretation on the time axis, states are true at moments,
because nothing happens, whereas activities require intervals in their interpretation, because
they necessarily imply a development over time (Vendler, 1957; Bennett & Partee, 1972;
6
Dowty, 1979). Eat an apple in (5a) is an accomplishment in Vendler’s classification, because
it describes a culminating process the truth of which can only be evaluated at the entire
interval. In contrast, Vendler qualifies verb phrases like reach the top (10c) as achievements,
because they describe instantaneous transitions that are verified at instants. Other
classifications besides Vendler’s have been proposed in the literature; they are discussed by
Filip (this volume).
3. Grammatical Aspect from a Cross-Linguistic Perspective
5.1.
Perfective/Imperfective Aspect: Observations from Russian
English is not necessarily a good starting point for a broad perspective on grammatical aspect,
so this section places the issue in a cross-linguistic perspective. As evidenced by typological
research (Comrie, 1976; Dahl, 1985), the distinction most commonly found in languages is
that between perfective and imperfective aspect (cf. Gvozdanovic, this volume). This
distinction is central to Slavic languages (section 3.1) and Romance languages (section 3.2),
but also many others, cf. Dahl and Velulipillai (2008). Section 3.3 returns to the English
Perfect and Progressive. Languages which lack the category of verbal tense, such as SinoTibetan languages, often have highly complex aspectual systems (section 3.4).
The perf(ective)/imp(erfective) contrast has been extensively investigated for Slavic
languages. In a Slavic language like Russian, every verb in the lexicon is labeled as perfective
or imperfective. Affixation or stem alternations create perfective verbs out of imperfective
roots (11a) and vice versa (11b):
(11)
a.
pisat’ ‘to write’ (imp), na-pisat’ ‘to write (something)’ (perf), pod-pisat’
sign’ (perf). [Russian]
b. dat’ ‘to give’ (perf), da-va-t’ ‘to give’ (imp).
7
‘to
c.
pod-pis-yva-t’ ‘to sign’ (imp).
When different prefixes combine with the same verbs, they give rise to a range of
meanings. Verbs that only differ in aspectual value such as pisat’ ‘write’ (imp) and na-pisat’
‘write’ (perf) are called aspectual pairs. In contrast, the verb pod-pisat’ not only differs in
aspectual value from pisat’, but also takes up a different lexical meaning, meaning ‘sign’. Verbs
that contain such lexical prefixes may take the same suffix that an inherently perfective verb
takes (cf. 11b) to create a secondary imperfective: (11c) is the imperfective counterpart of the
perfective verb pod-pisat’ ‘to sign’.
Inflected verbs are always marked as perfective or imperfective in Slavic languages,
so the role of grammatical aspect in these languages is pervasive. According to Comrie (1976,
p. 16), perfectivity “indicates the view of the situation as a single whole, without distinction
of the various separate phases that make up that situation.” Smith (1991/1997) takes the
Russian perfective to include initial and final endpoints. For (12a), both characterizations
imply that the opening of the window is completed:
(12)
a.
On ot-krylperf
okno.
[Russian]
he open.PAST.PERF window.ACC
‘He opened (the/a) window.’
b. On ot-krylperf
okno
*(za) dva časa.
he open.PAST.PERF window.ACC *(in) two hours
‘He opened the window in two hours/*two hours (long)
c.
Ja pro-čitalpperf
knigu.
I
book.ACC
THROUGH-read.PAST.PERF
entails Ja (bol’še)
ne čitajuiimp knigu.
I (anymore) not read.PRES
‘I read the book’ entails ‘I am not reading the book anymore’
8
book
Perfective verb phrases combine with in-adverbials rather than for-adverbials (12b),
and give rise to the entailment that completed events in the past are incompatible with the
same situation ongoing at the speech time (12c).
The imperfective aspect “pays essential attention to the internal structure of the
situation” (Comrie, 1976, p. 16) and abstracts away from initial and final points (Smith
1991/1997). Prominent interpretations of the imperfective are the plain stative reading (13a,
from Smith, 1991, p. 318), the ongoing process reading (13b, from Borik, 2002, p. 48) and the
habitual reading ((13c), from Berit Gehrke, p.c.):
(13)
a.
Vanja golodalimp
Vanya was starving.IMP.
b. Petja peresekalimp
etot kanal
kogda načalsja
Peter cross.IMP.PAST.SG.MASC. this channel when
PERF-begin-PST.SG.MASC
štorm
storm
‘Peter was crossing this channel when the storm began.’
c.
On odin iz šesti detej, i ego otec bylimp takim sil’nym, čto, kogda synov’ja
klaliimp na stol orech, on – raz pal’cem! – raskalyvalimp ego lučše, čem ščipcy
dlja orechov. (Hrabal, The Mermaid)
‘He is one of six children, and his father was so strong that, whenever the sons put
a nut on the table, he – once with a finger – cracked it better than a nutcracker.’
With the ongoing process reading, there is no entailment that the event was completed,
as we see in (14) (compare with 6 above):
(14)
Anja ubiralaimp kvartiru. →
/ Anja ubralaperf kvartiru.
9
Anja tidied.imp flat. →
/ Anja tidied.perf flat.
‘Anja was tidying the flat.’ does not entail ‘Anja tidied the flat.’
Aspect affects temporal structure, as in (15, from Smith, 1991, pp. 301, 303).
Perfective sentences in the context of a when-clause have a sequential interpretation (15a),
whereas the imperfective describes the situation as underway at the time of the adverbial
clause situation (15b):
(15)
a.
Kogda on vyšel iz komnaty, on posidelperf v parke
When he went out of the room, he sat-PERF for a while in the park.
b. Vanja pelimp v parke, kogda Nina pojavilasperf’.
Vanja sang-IMP in the park when Nina appeared-PERF.
Besides the range of meanings illustrated in (13), the Russian imperfective can also
have a general factual meaning, as in (16a) or report an annulled result, as in (16b) (examples
from Grønn, 2003, p. 25):
(16)
a.
Ja vaši očerki o Sibiri čitalimp, mne oni očen’ nravjatsjaimp.
I have read your essays on Siberia, I like them a lot.
b. A: Iskaliimp menja?
Have they been looking for me?
B: Milicioner nedavno priezžalimp. … S otcom razgovarivalimp.
A police officer arrived recently. … He talked to your father.
10
The imperfective verb čitalimp in (16a) refers to a complete single, telic event in the
past. Under the factive interpretation, the kind of entailment in (14) does go through (cf.
Borik, 2002). From the use of the imperfective priezžalimp in (16b) we can infer that the police
officer has left at the time of the conversation. The use of priechalperf in the context of (16b)
would imply that the police were still present.
The perfective/imperfective is a morphological distinction also found in the present
tense. However, perfective interpretations are incompatible with present time reference, and
in many languages, including Russian, the perfective non-past tense has future reference (17,
Dahl, 1985, p. 80):
(17)
Ja napišuperf
pis’mo.
I perf.write.non-past
letter
‘I will write a/the letter.’
The special meanings of the Russian imperfective in (16) are not necessarily shared by
all languages with a perfective/imperfective distinction, but we find counterparts to the
observations in (12)-(15) as well as (17) across a wide range of languages.
5.2.
Perfective/Imperfective Contrast: Observations from French
In Romance languages, the perfective/imperfective contrast is confined to the past tense,
where tense and aspect are morphologically fused. Thus, the Italian Passato Remoto feci ‘I
did’ (perfective past) is aspectually opposed to the Imperfetto facevo ‘I was doing/used to do’
(imperfective past) (cf. Bertinetto & Delfitto, 2000). Similar observations can be made for the
Spanish Préterito escribió ‘He wrote’ (perfective past) vs. the Imperfecto escribía ‘he was
writing/used to write’ (imperfective past) (cf. Cipria & Roberts, 2000) or the French Passé
11
Simple il écrivit ‘he wrote’ (perfective past) vs. the Imparfait il écrivait ‘he wrote’
(imperfective past).
The aspectual contrast between the French Passé Simple (PS) in (18a) and the
Imparfait (Imp) in (18b) resembles that between the Simple Past and the Progressive in (2a)
and (b):
(18)
a.
Il écrivit
sa thèse en 2009.
[French]
He wrote.PS his thesis in 2009.
‘He wrote his thesis in 2009.’
Il l’a fini en Septembre.
He finished in September.
#Je pense qu’il est en train d’écrire la conclusion maintenant.
#I think he is writing the conclusion now.
#Il n’a jamais fini, car il est mort en Septembre.
#He never finished, for he died in September.
b. Il écrivait
sa
thèse en 2009.
He wrote.IMP his thesis in 2009
‘He was writing his thesis in 2009.’
Il a fini en Septembre.
He finished in September.
Je pense qu’il est en train d’écrire la conclusion maintenant.
I think he is writing the conclusion now.
Il n’a jamais fini, car il est mort en Septembre.
He never finished, for he died in September.
12
With the ongoing process reading, there is no entailment that the event culminates,
that is, (18b) does not entail (18a) (cf. also 6b and 14 above). However, the Imparfait is not
incompatible with a statement affirming completion (cf. 2b). The Passé Simple does not allow
the process to continue at the speech time, whereas the Imparfait does.
The Passé Simple and Imparfait affect temporal structure in a similar way as we
observed in (2) and (15). (19a) illustrates temporal overlap with the Imparfait, and (19b)
succession in time with the Passé Simple (examples from Molendijk, 1990):
(19)
a.
Quand l’ennemi attaqua son armée, le général se retirait.
When the enemy attacked his army, the general witdrew.imp
‘When the enemy attacked his army, the general was withdrawing.’
b. Quand l’ennemi attaquait son armée, le général se retira.
When the enemy attacked his army, the general withdrew.ps
‘When the enemy attacked his army, the general withdrew.’
The Slavic and Romance imperfective are often compared to the English Progressive.
Note though that Romance languages have special periphrastic constructions for progressive
means, illustrated in (20):
(20)
a.
Yo estaba hablando con mi madre
[Spanish]
‘I was talking with my mother.’
b. Nous sommes en train de chercher un nouvel appartement.
‘We are looking for a new apartment.’
13
[French]
Just like the Slavic imperfective (cf. 13a), the Romance imperfective is compatible
with stative verbs (21a), and allows a habitual interpretation (21b) besides the ongoing
process reading in (19a), compare (13c):
(21)
a.
Julie était
amoureuse
Julie was.imp
in.love
de Marc
of Marc
‘Julie was in love with Marc.’
b. Le samedi,
Julie jouait
au tennis.
The Saturday, Julie played.imp at tennis
‘On Saturdays, Julie played tennis.’
Just like the Russian imperfective (cf. 16a), the French Imparfait can refer to a
completed event, as pointed out by Brunot and Bruneau (1949) who cite examples like (22)
(translation by Binnick, 1991, p. 374):
(22)
Lorsque le notaire arriva avec M. Geoffrin … elle les reçut elle-même et les invita à
tout visiter en détail. Un mois plus tard, elle signait le contrat de vente et achetait en
même temps une petite maison bourgeoise.
(Maupassant, Une vie, p. 292)
‘When the notary arrived.PS with M. Geoffrin … she received.PS them herself, and
invited.PS them to examine everything in detail. A month later, she signed.IMP the
contract of sale and bought.IMP [at the same time] a little bourgeois house.’
The three events of arriving, receiving and inviting are reported as complete events in
the Passé Simple. Signing the contract and buying a house are related in the Imparfait, but
clearly refer to completed events in the past. This use of the Imparfait depends on adverbials
14
such as un mois plus tard ‘a month later’ in (22) moving the story time, which has given rise
to the term “Imparfait de rupture” (Tasmowski, 1985), and emphasizes the discourse function
of aspect (cf. Brès, 2005, and references therein; and Carruthers, this volume).
The range of meanings found for the Romance perfective/imperfective past resembles
the Slavic situation. Yet, there are differences, relating to the interaction between grammatical
aspect and situation class (in Slavic) and the interaction between grammatical aspect and tense
(in Romance). These compositionality issues are addressed in section 4.
5.3.
Progressive and Perfect: Observations from English
The perfective/imperfective contrast is often viewed as the typologically most prominent
aspectual distinction, but not all languages fit this binary approach. English doesn’t, for
instance, for it has a Progressive, which only grammaticalizes a subpart of the meaning of the
imperfective. The focus of the Progressive on ongoing processes typically restricts its
application to non-stative verbs (cf. 8 above). Unlike the Slavic and Romance Imperfective,
the English Progressive does not have a habitual interpretation. Unlike the Simple Past in
(2a), the Past Progressive in (2b) does not commit the speaker to the claim that the
culmination point (completion of the dissertation) is ever reached, a reading referred to as the
“imperfective paradox.” Accordingly, the Progressive is generally assigned a modal
semantics, following Dowty (1979). Under the modal analysis, a Progressive sentence
requires part of the process to be realized in the actual world, but leaves open the remaining
part including its culmination point, or situates this part of the process in some possible world,
with strict requirements on accessibility from the actual world (compare Vlach, 1981;
Parsons, 1990; Landman, 1992; Portner, 1998; and Zucchi, 1999, for a range of proposals,
and Mair, this volume, for an overview).
15
The Simple Past in (2a) is sometimes described as perfective (e.g. Smith, 1991/1997),
because it presents the writing of the dissertation as a completed event. Alternatively, we can
view the Simple Past as an aspectually neutral tense, which just locates the state or event
introduced by the predicate-argument structure in the past (Kamp & Reyle, 1993; de Swart,
1998). Given that ‘write a dissertation’ is an accomplishment, both approaches imply that the
event in (2a) culminates before the speech time. The neutral interpretation accounts more
easily for stative descriptions (1), or the habitual interpretation of the Simple Past (23):
(23) Julia played tennis on Saturdays.
Besides the Progressive, English also has a perfect/non-perfect distinction, which is
viewed as a temporal operator by Reichenbach (1947) and Verkuyl (1999), but included in the
category of aspect by Comrie (1976); see Ritz, this volume. Across languages, the
perfect/non-perfect contrast is generally expressed periphrastically (Dahl and Velupillai,
2008).
The Perfect is not to be confused with the Perfective. Slavic languages grammaticalize
the perfective/non-perfective contrast, but do not have a perfect. Romance languages on the
other hand grammaticalize both a perfective/imperfective contrast in the past tense (French
Passé Simple donna ‘gave.PERF’ vs, Imparfait donnait ‘gave.IMP’) and a perfect/non-perfect
contrast in all tenses (French Passé Composé a donné ‘has given’, Plus-que-Parfait avait
donné ‘had given’).
Reichenbach (1947) used the contrast between (24a) and (b) to motivate a
representation that relies on three points, namely S (for speech time), E (for event time) and R
(for reference time):
16
(24)
a.
Julia left the party.
E,R – S
b. Julia has left the party.
E – R,S
In both (24a) and (b), the event of Julia leaving the party is situated before the speech
time, so two points do not suffice to distinguish the temporal structure of the two. The
structure E,R – S for the Simple Past indicates that the event in (24a) is viewed from a
reference point coinciding with the event, whereas the reference point for the event in (24b)
coincides with the speech time (E – R,S). The deictic character of the Present Perfect in (24b)
blocks the use of time adverbials locating the event at a particular point in time:
(25)
a.
Julia left the party at 10pm.
b. #Julia has left the party at 10pm.
However, the observations in (25) hold for the Present Perfect in standard British
English, but not necessarily in other varieties of the language (cf. Ritz, in press, and this
volume), and they do not hold for counterparts of the Present Perfect in languages like Dutch,
German or French (cf. de Swart, 2007).
Comrie (1976), Moens and Steedman (1988), Kamp and Reyle (1993), and others
offer aspectual analyses of the perfect in terms of mapping of the event into a post-state of the
event (result state or otherwise). As Comrie (1976, p. 60) points out, perfects vary across
languages in whether they do (e.g. English) or do not (e.g. Dutch) allow a so-called
continuative interpretation. The continuative interpretation of (26a) locates the initial
boundary of the state of living in the past, and focuses on the result of starting the state, which
extends into the present (interpretation 26ai):
17
(26)
a.
Mary has lived in London for five years.
(i) Mary moved to London five years ago, and still lives there.
(ii) Mary lived in London during a five year period in the past.
b. Maria heeft vijf jaar
in Londen gewoond.
[Dutch]
Maria has five years in London lived.
The continuative perfect reading is available in English (26ai), but not in its Dutch
counterpart (26b), which only allows the reading in (26aii).
De Swart (2007) and Ritz (in press) emphasize that we should combine temporal and
aspectual features of the perfect in order to explain its discourse behaviour. The
Reichenbachian analysis suggests that perfects do not have a narrative use, for their reference
point coincides with the speech time. The (British) English present perfect (PP) is indeed
blocked in narrative when-clauses (27a), where we have to use the Simple Past (SP), but its
French counterpart the Passé Composé (PC) is perfectly felicitous in this environment (27b):
(27)
a.
*When John has seen (PP) me, he has got (PP)/got (SP) frightened.
b. Quand Jean m’a vu (PC), il a eu peur (PC).
[French]
Portner (2003) provides a recent overview of proposals, and a distinction between
semantic and pragmatic features of the interpretation of the perfect, where the primary
semantic contribution of the perfect is described as temporal in nature. See also Ritz, this
volume.
Compare Schmitt (2001) and Schaden (2009) for more observations on the perfect in a
cross-linguistic perspective.
The Perfect and Progressive can be combined in sentences like (28):
18
(28)
a.
Professor P. is the head of the writing department, where he has been writing
and performing since 2002.
b. [the dwarves in the story of Snow White and the seven dwarves:]
Who has been sitting in my chair? Who has been eating from my plate?
Who has been drinking from my cup?
These combinations allow both continuative (28a) and non-continuative (28b)
interpretations. Their interpretation requires a handle on the compositional semantics of tense
and aspect markers, which will be addressed in section 4.
5.4.
Multiple Aspectual Distinctions: Observations from Mandarin Chinese
Besides English, many other languages display rich aspectual systems that are difficult to
relate to the perfective/imperfective or the perfect/non-perfect distinction. Here I offer some
observations from Mandarin Chinese to illustrate. Sino-Tibetan languages are well known for
lacking the category of verbal tense (cf. Lin, this volume). They make up for it by a rich
aspectual system. Mandarin Chinese is strongly isolating, so these aspectual distinctions are
not rendered by inflection on the verb, but by separate markers in the sentence. Sentences
without aspect markers can often be used with different temporal interpretations, as illustrated
in (29a) and (b) (from Lin, 2003b):
(29)
a.
Lisi hen jushang.
Lisi very depressed
‘Lisi is very depressed.’
b. Zhangsan dapuo yi-ge huaping
19
Zhangsan break one-cl vase
‘Zhangsan broke a vase.’
Lin (2003b) and Smith and Erbaugh (2005) point out that in isolation, sentences
without time adverbials or aspect markers, describing atelic situations, tend to get a present
tense interpretation (29a), but those describing telic situation get past time reference (29b),
along the lines of Bohnemeyer and Swift’s analysis of default aspect (2004).
Mandarin has a range of aspectual particles that may affect location in time, and
corpus studies indicate that such markers are widely used (Xiao & McEnery, 2004). Example
(30a) (from Smith, 1991, p. 349) illustrates that sentences marked with the perfective particle
le often describe completed events in the past. However, as (30b, from Smith, 1991, p. 349)
illustrates, le merely requires boundedness of the event, not necessarily completion, which is
conveyed by the resultative suffix -wan ‘finish’ (compare also Soh & Kuo, 2005).
(30)
a.
Wo shuiaduan le
I
break
tui
ASP leg
‘I broke my leg (it’s still in a cast).’
b. Wo zuotian
I
xie
le
gei
yesterday write asp to
Zhangsan de xin,
keshi mei xie - wan.
Zhangsan de letter,
but not write-finish
‘I wrote a letter to Zhangsan yesterday, but I didn’t finish it.’
c.
*Liming ai-le
Xiaojuan.
Liming
Xiaojuan
love-ASP
‘Liming loved Xiaojuan.’
d. Liming ai-le
Xiaojuan san-nian.
Liming love-ASP Xiaojuan three-year
20
‘Liming loved Xiaojuan for three years.’
e.
Tamen daoda shand-ding
They
le.
reach mountain-top asp
‘They reached the top of the mountain.’
Le is incompatible with stative verbs (30c), unless they are bounded by a for-adverbial
(30d) (Xiao & McEnery, 2004). There is a difference between verbal -le, as illustrated in
(30a-d) and sentence-final -le (30e, from Soh, 2009). Verbal -le is generally treated as a
perfective marker, whereas the sentence-final -le presupposes a transition and is closer to a
perfect.
Other aspectual markers include the experiential particle guo in (31a from Smith,
1991, p. 349), which can be combined with -le, as illustrated in (31b) (from Lin, 2003b):
(31)
a.
Wo shuaiduan guo tui
I break
asp leg
‘I broke my leg (it has healed since).’
b. Wo chi guo le.
I
eat ASP ASP
‘I have eaten.’
c.
Lisi zai xi-zao
Lisi ASP take-bath
‘Lisi is taking a bath.’
d. Ta zui li jiao-zhe koxiangtang
he mouth inside chew-Asp chewing-gum
‘He is chewing a chewing gum in his mouth.’
21
The experiential particle guo in (31a) requires a discontinuity with the speech time that
le in (30a) (and the English Perfect) lack. Zai and zhe in (31c,d, from Lin, 2003b) focus on the
internal development of the situation. Just like the English Progressive, zai (31c) is restricted
to dynamic (non-stative) verbs. Zhe (31d) only combines with atelic verb phrases. The
examples in (30) and (31) give a mere indication of the complexity of verbal aspect in
Mandarin Chinese.
The data provided here for Russian, French, English and Mandarin Chinese only
scratch the surface of the complexity of verbal aspect in these languages. However, the
various ways of establishing aspectual distinctions in the grammar can be compared in terms
of the readings they have, the entailments and continuations they permit, the effects they have
on temporal structure, and the distribution of labour between the members of the temporalaspectual system of the language. A full-fledged theory of grammatical aspect is very
complex, because it needs to take into account the cross-linguistic variations and the
interactions of aspectual markers with aspectual class and with the rest of the grammar.
4. The Compositional Structure of Aspect
5.1.
A Layered Representation
The distinctions established between aspectual class, grammatical aspect and tense raise the
question how these notions interact. In this section, I assume a layered structure of aspect, and
discuss a range of challenges for the compositional interpretation of aspect. Section 4.1
introduces the basic ideas, section 4.2 focuses on the relation between perfectivity and telicity,
and section 4.3 on the amalgamation of tense and grammatical aspect.
22
Many current theories adopt some version of a layered representation in which tense
syntactically and/or semantically dominates grammatical aspect, which in turn dominates
aspectual class.
(32) [Tense [Aspect* [aspectual class]]]
The Kleene star * is well known from mathematical logic and indicates that aspectual
markers can occur 0, 1, 2 …n times in the structure. (33) spells out the layered structure of
(1a), (2a) and (2b):
(33)
a.
Bill was in love with Susan.
[PAST [ Bill in love with Susan ]]
b. Sarah wrote a dissertation.
[ PAST [ Sarah write a dissertation ]]
c.
Sarah was writing a dissertation.
[ PAST [ PROG [ Sarah write a dissertation ]]]
The difference in aspectual nature is located in the presence of a grammatical marker
PROG
in (33c), versus its absence in (33a) and (33b).
There is substantial disagreement in the literature about the way such a layered
aspectual structure should be set up and interpreted. Maybe tense is not projected in tenseless
languages such as Mandarin Chinese. Some languages are argued to have ‘lower’ or ‘higher’
aspectual projections than others, giving rise to a range of possible functional structures
within the verb phrase (cf. Verkuyl, 1999; Travis, 2000, 2010; Ritter & Rosen, 2005;
Ramchand, 2008).
(For example, the progressive can be viewed as a VP operator, and the
23
tense as an operator at the sentential level, as in (39b) below, for instance. This chapter will
not be concerned with the difference between VP and S-level, which becomes relevant,
however, in sentences that contain scope-bearing operators in subject position.)
Some researchers defend a two-component theory in which grammatical aspect and
aspectual class are interpreted by means of different sets of tools (Smith, 1991/1997;
Depraetere, 1995; Filip, 1999; Bertinetto & Delfitto, 2000), others use the same semantic
machinery for both, whichever that may be (Moens & Steedman, 1988; Parsons, 1990; Kamp
& Reyle, 1993; de Swart, 1998; Verkuyl, 1999; Cipria & Roberts, 2000). What both lines of
analysis agree on is that grammatical aspect determines the aspectual nature of the sentence as
a whole, and may overrule certain semantic features of its internal aspectual make-up.
The interactions between aspectual class and grammatical aspect complicate the
debate. For English, we have already seen that the Progressive is normally restricted to nonstative verbs (2b vs. 8a). This strongly suggests that the interpretation of the Progressive is
closely intertwined with the semantics of the verb it applies to, instead of being fully
independent from it. Similar observations can be made for Mandarin Chinese, where the
progressive marker zai is restricted to non-stative verbs (cf. section 3.4 above). Whether such
dependency relations are better modeled as constraints on transitions in an aspectual network
(Moens & Steedman, 1988) or as marked aspectual choices in a two-component theory
(Smith, 1991, p. 226) remains an open issue.
The layered structure in (32) helps us analyze sentences such as (34a) and (b), which
contain multiple aspect markers in terms of the recursive application of aspectual operators:
(34)
a.
Who has been sitting on my chair?
?x [PRES [ PERF [ PROG [ x sitting on speaker’s chair]]]]
b. Ta he-le
san-wan
tang le.
[Mandarin Chinese]
24
‘He drank three bowls of soup.’
[ LE [ LE [he drink three-bowl soup]]]
c.
pod-pis-yva-t’ ‘to sign’ (imp)
[Russian]
[IMP [ PERF [write]]]
The Perfect in (34a) takes wide scope over the Progressive, and the inverse scope
reading is not available. Soh and Gao (2006) argue that in double -le sentences in Mandarin,
sentence-final -le necessarily takes wide scope over verbal -le. Similarly, the imperfective
suffix -(y)va- takes wide scope over the lexical perfective prefix pod- in the Russian
secondary imperfective construction in (34c). So the iteration of aspectual markers is subject
to ordering constraints in the semantics. Koenig and Muansuwan (2005) carry out an
extensive study of recursive applications of aspect markers in Thai.
Although there are interactions between aspectual class and grammatical aspect in
languages like English, we can still tease the two apart in a representation like (33). This is
much more difficult in languages in which the boundaries between the two aspectual
categories are not strict, such as Russian and Mandarin Chinese (section 4.2). Once we realize
that grammatical markers can impose structure on a category they dominate, we can also
address such effects in the relation between tense and grammatical aspect (4.3), as well as
grammatical aspect and other aspectually sensitive expressions (section 5).
5.2.
Perfectivity and Telicity in Russian and Mandarin Chinese
Given that most Slavic languages do not have articles, the contrast between atelic verb
phrases like read articles and telic verb phrases like read the articles is not made overtly.
Russian bare plurals and bare mass nouns are interpreted as definite or ‘specific’ in the
context of a perfective verb form (35b), whereas they can be definite or indefinite in the
25
context of an imperfective verb form (35a) (examples from Borik, 2002, p. 68, based on
insights going back to Schoorlemmer, 1995). Thus the perfective verb is necessarily telic:
(35)
a.
Petja čitalimp
stat’i/literaturu
Peter read-IMP-PAST.SG.MASC. articles/literature-ACC
‘Peter was reading articles/the articles/literature/the literature/read
articles/literature’
b. Petja pro-čitalperf
stat’i/literaturu
Peter PERF-read-PAST.SG.MASC articles/literature-ACC
‘Peter read the articles/the literature’
This intertwining of grammatical aspect and aspectual class has given rise to a debate
on the status of prefixes in Slavic languages: do they count as perfectivity markers, as argued
by Smith (1991/1997) and Borik (2002), or are they telicity inducing, as claimed by Filip
(1999) (for Czech) and Arsenijvič (2006) (for Serbo-Croatian)? Di Sciullo and Slabakova
(2005), Gehrke (2007), and Ramchand (2008) suggest that the way out of this dilemma is to
establish a distinction between two kinds of prefixes.
Lexical affixes that create a different verb meaning qualify as internal affixes which
affect the argument structure and bear on telicity. For instance, when the imperfective root
pisat’ ‘to write’ combines with the lexical prefix pod-, the result is a perfective verb podpisat’ with a different lexical meaning ‘to sign’ (cf. 11 above). Prefixes such as na- in (11a)
which build aspectual pairs also qualify as internal prefixes. External prefixes such as po-in
po-spall’ ‘sleep for a while’ or za- in za-pisat’ ‘to begin to write’ bear on the event as a
whole. Gehrke (2007, p. 171) shows that internal prefixes always induce telicity (36a). In
26
contrast, externally prefixed predicates are not necessarily telic, as illustrated in (36b) (cf. also
Filip, 2000):
(36)
a.
Ja na-pisalperf pis’mo
*(za) dve minuty.
I perf.wrote letter.acc *(in) two minutes
‘I wrote a/the letter in/*for two minutes.’
b. On po-spalperf (*za) dve minuty.
He perf.slept (*in) two minutes
‘He slept *in/for two minutes.’
Under the approach advocated by Gehrke and others, internal prefixes contribute to
aspectual class, whereas external prefixes contribute to grammatical aspect. The distinction
cannot be drawn on morphological grounds, because both markers are part of the inflectional
system of the verb. Note that verbs can take two perfective prefixes, but their order is
constrained: kopit' ‘to save-IMP’, na-kopit' ‘PERF-to save up’ and pod-na-kopit' ‘PERF-PERF-to
save up some’, but *na-pod-kopit' (Borik, 2002). According to Gehrke (2007, p. 170),
external prefixes can be stacked on top of internal prefixes, but not the other way around.
The close connection between aspectual class and grammatical class in Russian is
reason for caution in the compositional analysis of aspect in a cross-linguistic perspective. We
also find this phenomenon in Mandarin Chinese. Mandarin Chinese does not have articles,
and also lacks the category of number, so there is no formal distinction between singular and
plural nouns. As a result, a bare nominal such as N could mean ‘a N’, ‘the N’, ‘N’s’ or ‘the
Ns’. This has similar consequences for aspectual distinctions grounded in predicate-argument
structure as what we saw for Russian. Sybesma (1999) notes that bare nominals get a definite
27
or specific interpretation with a range of predicates that are bounded for reasons independent
of the object, as in (37):
(37)
a.
Wo he-guan le tang.
I
[Mandarin Chinese]
drink-up ASP soup
‘I finished the/*some soup.’
b. Wo mai-zhao le shu.
I
buy-get asp book
‘I managed to buy the/*some books.’
In section 1.1, we observed that English has inherently telic verbs (eat up, drink up),
but in Mandarin we construct such verbs. Sybesma defines the predicates in (37a, b) as
perfective predicates, which have a built-in moment of completion.
In both Russian and Mandarin Chinese we observe that perfectivity imposes telicity in
the absence of articles. Other aspectual markers (imperfective in Russian, guo etc. in
Mandarin) do not have such effects, so grammatical aspect and aspectual class are easier to
tease apart in those constructions. However, in a language like Finnish, partitive case may
obliterate the distinction between imperfectivity and atelicity (cf. Kiparsky, 1998, and
Richardson, this volume).
5.3.
Amalgamating Tense and Grammatical Aspect in French
Further complexities arise in Romance languages, in which the past tense and perfective/
imperfective aspect are morphologically fused (cf. section 3.2). For examples like (18a),
repeated here as (38a), a separation of temporal and aspectual information in the layered
structure would seemingly lead to the representation in (38b).
28
(38)
a.
Il écrivit
sa thèse en 2009. [French]
He wrote.ps his thesis in 2009.
‘He wrote his thesis in 2009.’
b. [ past [ perf [ he write his thesis ]]]
c.
[ past [ he write his thesis ]]
The problem with (38b) is that the morphology of the Passé Simple does not
compositionally map onto this structure: the verb form écrivit cannot be split up into a part
that leads to the past tense operator, and a part that introduces the perfective operator. As an
alternative, de Swart (1998) adopts the structure in (38c). The Passé Simple introduces just a
past tense operator, but requires the predicate-argument structure to introduce an event. Given
that ‘he write his thesis’ qualifies as an accomplishment (cf. section 1.2), (38c) locates a
bounded event in the past just like the English (2a) does, which does not carry overt perfective
morphology.
Under de Swart’s (1998) analysis, the Imparfait gets the same semantics as the Passé
Simple: it introduces a past tense operator. However, whereas the Passé Simple locates an
event in the past, the Imparfait requires the predicate-argument description to contribute a
state or an unbounded process. An example like (21a), repeated here as (39a), can then be
analyzed as in (39b):
(39)
a.
Julie était
amoureuse
de Marc.
Julie was.IMP
in.love
of Marc
‘Julie was in love with Marc.’
b.
[PAST[Julie is in love with Marc]]
29
The Imparfait sentence in (39a) locates the state of Julie being in love with Marc in the
period preceding the speech time. The semantics of (39a) is then similar to that of the English
example (33a) above, which does not carry overt imperfective morphology.
5.4.
Aspectual Coercion
In line with this analysis, corpus research indicates that the Romance perfective past tense is
found predominantly with event predicates, and the imperfective past is found mostly with
state and activity predicates (cf. de Jonge, 2000 for Spanish). But of course, both past tense
forms are grammatically possible with any type of eventuality. Under de Swart’s (1998)
analysis, the combination of the Passé Simple with a state/activity verb, or the combination of
the Imparfait with an event predicate requires a process of aspectual coercion. Coercion is the
process by means of which an argument adapts to the requirement of the functor with which it
combines (cf. Pustejovsky, 1995). Aspectual coercion then requires the eventuality
description to shift its meaning to satisfy the aspectual selection requirements of the Passé
Simple or Imparfait. For an Imparfait sentence like (18b), repeated here as (40a), this leads to
the representation in (40b), where Ceh indicates the coercion of an event predicate into a
homogeneous description.
(40)
a.
Il écrivait
sa thèse en 2009.
He wrote.IMP his thesis in 2009
‘He was writing his thesis in 2009.’
b.
[ PAST [Ceh [[ he write his thesis ]]]
30
The coercion operator Ceh is located in the grammatical aspect slot of the layered
representation, but is not overt. In this case, the ongoing process reading is the most likely
interpretation of Ceh, but for an example like (21b) above, Ceh leads to a habitual
interpretation. Although the coercion operator allows for a range of interpretations, it cannot
be semantically empty: there must be an aspectual shift in meaning associated with it. The
interpretation of Ceh is always associated with a shift from events to states or unbounded
processes. The coercion operator Che, which comes into play when the Passé Simple combines
with a state or an event, often triggers an inchoative reading as in (41a) (from Molendijk,
1990, p. 93):
(41)
a.
Jean
inventa
une machine à
traduire. Il
connut
la
gloire.
Jean
invented.PS a
machine to translate. He knew.PS the glory.
‘Jean invented a translation machine. He received praise.’
b.
[ PAST [ Che [ he knew the glory ]]
The analysis in terms of aspectual coercion is not uncontroversial. Bonami (2002) and
Caudal (2005) propose implicit operators for the Imparfait, which are lexically licensed, and
thus emphasize the role of aspectual class. Cipria and Roberts (2000) argue for Spanish that
the different readings of the Imperfecto are inherent to the truth-conditional semantics of this
tense form. Bonomi (1997) adopts a comparable approach to Italian, cf. also Bertinetto and
Delfitto (2000). One way of settling the debate on aspectual coercion vs. underspecificity or
ambiguity is to rely on evidence from processing. However, these experimental data remain
difficult to interpret (cf. de Swart, to appear).
31
Cross-linguistically, we find similar effects of aspectual reinterpretation in Russian
and Mandarin. According to Smith (1991/1997), stative verbs do not take perfective prefixes
in Russian. The rare instances of perfective states give rise to inchoative readings as in
ponjat’, the perfective of ponimat’ ‘understand’ in (42a, from Comrie, 1976, p. 19). Soh
(2009) provides similar examples for Mandarin Chinese (42b):
(42)
a.
Nakonec on ponjal, v čem delo.
[Russian]
‘At last he grasped what was up.’
b. Ta zhidao-le zhe-jian shi.
[Mandarin]
he know-asp this-cl matter
‘He came to know this matter.’
States are normally not compatible with the aspectual particle -le (cf. section 3.4
above), so Soh (2009) takes (42b) to describe an achievement, suggesting a reinterpretation of
the stative verb.
Effects of aspectual coercion also arise with English for- and in-adverbials that are
sensitive to the telic/atelic nature of the predicate-argument structure they apply to. As
pointed out in section 2.1 above, sentences containing an event predicate are not always
infelicitous with a for- adverbial, but give rise to a special interpretation:
(43)
a.
Jim hit a golf ball into the lake for an hour.
[ PAST [ for an hour [ Ceh [ Jim hit a golf ball into the lake ]]]]
b. The baby was asleep in ten minutes.
[ past [ in ten minutes [ Che [the baby be asleep]]]]
32
(43a) gives rise to an iterative interpretation such that the same golf ball repeatedly
ends up in the lake (Van Geenhoven, 2005). The aspectual reinterpretation of the predicateargument structure is located in the coercion operator Ceh, mapping the event onto an atelic
(iterative) situation that can be measured out by the for-adverbial. The inchoative
interpretation of was in (43b) implies that the initial point of the baby being asleep was ten
minutes after the moment where the time measurement started (e.g. the moment she was put
to bed). Given that the in-adverbial requires an event predicate, the atelic predicate be is
reinterpretated as a transition between two states by the coercion operator Che.
The analysis spelled out in (43) broadens the scope of the layered structure in (32) to
include other aspectually sensitive expressions besides grammatical aspect markers. We
further investigate these in section 5.
5. Grammatical Aspect and Other Aspect-Sensitive Expressions
Section 4 ended by extending the layered structure in (40) to account for the interaction of
grammatical aspect with aspectual adverbials like for an hour/in an hour. In this section, we
broaden the approach to a cross-linguistic investigation of the aspectual requirements of
measurement adverbials (5.1), and to the interaction of aspect with the marker of sentential
negation not and frequency adverbs like always, often (5.2).
5.1.
Interaction of Grammatical Aspect with Measurement Adverbials
As we saw in Section 4.2, the amalgamation between grammatical aspect and aspectual class
makes it difficult to determine whether Russian in- and for-adverbials are sensitive to
perfectivity or telicity. The examples in (36), repeated here as (44) have been used to argue in
favor of a distinction between internal and external affixes. Under the layered structure of
aspect, they can be handled in terms of scope:
33
(44)
a.
Ja na-pisalperf pis’mo
I
PERF.wrote
*(za) dve minuty.
letter.ACC *(in) two minutes
‘I wrote a/the letter in/*for two minutes.’
[ PAST [ in two minutes [ PERF [ I write a letter ]]]]
b. On po-spalperf (*za) dve minuty.
He perf.slept (*in) two minutes
‘He slept *in/for two minutes.’
[ past [ perf [ for two minutes [ he sleep ]]]]
Under the assumption that the internal affix in (44a) is located in a ‘low’ aspectual
position, it takes scope under the measurement adverbial, leading to an interpretation in which
the letter took two minutes to complete. The for-adverbial is not felicitous here, because the
aspectual adverbial applies to a perfective verb, which is not the right input. If the external
affix in (44b) is located in a ‘high’ aspectual position, it scopes over the duration adverbial.
Perfective aspect here marks the boundedness of the situation of sleeping for two minutes.
The in-adverbial is not felicitous here, because sleep is an activity verb. The interpretation of
measurement expressions is thus in line with the distinction between internal and external
prefixes.
As de Swart (1998) points out, both for- and in-adverbials require the Passé Simple in
French, which suggests that the adverbial scopes below the aspectually sensitive tense marker:
(45)
a.
La
rebellion
fit
rage
pendant
six
ans.
The
revolt
made.PS
rage
for
six
years
‘The revolt raged for six years.’
34
[ PAST [ for six years [ the revolt rage ]]]
b. On
finit
le
One finished.ps the
débat en
trente minutes.
debate in
thirty minutes.
‘The debate was finished in thirty minutes.’
[ past [ in thirty minutes [ finish the debate ]]]
The French Passé Simple amalgamates past time reference and perfective aspect (cf.
section 4.3 above). Given that the tense marker always takes widest scope in the layered
structure in (32), and both in- and for-adverbials lead to a bounded situation with well-defined
initial and final points, both structures in (45a) and (b) satisfy the aspectual selections of the
Passé Simple rather than the Imparfait.
More complex situations arise when the measurement phrase interacts with multiple
aspectual markers, as in the Spanish example (46) (from de Swart, 1998):
(46)
Toda la tarde
all
estuvieron entrando visitas.
the afternoon were.PRÉT
[Spanish]
coming visitors
[ PASTperf [ all afternoon [ PROG [ visitors come ]]]]
The situation in (46) is presented as a single complete whole (hence the Préterito,
PRET),
but consists of an ongoing action (hence the Progressive). As an aspectually sensitive
tense, we expect the Spanish Préterito to take wide scope over the Progressive as well as the
for-adverbial. Given that toda la tarde measures the duration of the ongoing situation, the
adverbial scopes over the Progressive.
The observations made with respect to (44)-(46) suggest that the high or low position
of aspect in the language determines the interaction of grammatical aspect with aspectually
35
sensitive expressions such as measurement phrases. Section 5.2 extends this view to negation
and frequency adverbials.
5.2.
Recursive Applications of Asp: Negation and Frequency
Negation takes scope over the Progressive in English (47a). The Perfect allows for scope
ambiguities (47b and c, from de Swart & Molendijk, 1999):
(47)
a.
Anne was not swimming across the Channel.
[ PAST [ NEG [PROG [Anne-swim-across-the-Channel]]]]
b. Mary hasn’t met the president.
[ past [ neg [perf [Anne meet the president]]]]
c.
Mary hasn’t written a single poem (for twelve years now).
[ pres [ perf [ neg [ Mary write a poem]]]]
There is no result state of meeting in (47b), and the state of not-writing started 12
years ago (continuative perfect, 47c).
Negation also takes wide scope over perfective verbs in Russian, compare (48a) and
(b) from Smith (1991, p. 335):
(48)
a.
Ja ešče ne čitalimp ètu stat’ju.
I yet not read-IMP this article
‘I have not read this article.’ (no reading)
b. Ja ešče ne pročitalperf ètu stat’ju.
I yet not read-perf this article.
‘I have not finished reading this article.’ (no completion)
36
As Smith explains, the perfective (48b) denies that the speaker has completed reading
the article, whereas the imperfective (48a) denies that the action was initiated at all.
No such effect is found in Romance languages, where negated sentences generally
combine with the Imperfective past tense form. Compare the French examples in (49a) (from
de Swart & Molendijk, 1999) and (49b):
(49)
a.
Jean
courait
Jean
ran.IMP
après Pauline.
after Pauline.
He
Il
NEG
ne
l’attrapait
her caught.IMP
pas.
NEG.
‘Jean ran after Pauline. He didn’t catch her.’
[ PAST [ NEG [ Jean catch Pauline ]]]
b. Les
The
jeunes gens
young
ne vinrent pas
chez le vieil homme
people NEG came.PS NEG by the old
quelque
temps.
some
time
man
pendant
for
‘The young people didn’t visit the old man for a while.’
[ past [ for some time [ neg [ the young people visited the old man ]]]]
De Swart and Molendijk attribute the preference of negation for combining with the
Imparfait to the fact that the aspectual distinction between perfective and imperfective is fused
with the past tense, which takes wide scope over negation, locating the absence of action in
the past. Given that absence of action normally counts as a state, the aspectual restrictions of
the Imparfait are satisfied in (49a). As illustrated in (49b), the Passé Simple is possible when
37
the negative situation is bounded by a for-adverbial, which we expect under the analysis
developed in section 5.1.
Similarly, the negation marker bu in Mandarin Chinese is incompatible with the
perfective marker -le, as illustrated in (50a) and (b) (from Ernst, 1995):
(50)
a.
Wo bu chi mugua
I
[Mandarin Chinese]
not eat papaya
‘I do not eat papaya.’
b. *Wo bu chi-le papaya.
c.
Wo mei-you chi mugua.
I
not-perf eat papaya
‘I didn’t eat papaya.’
But here the explanation is somewhat different. According to Ernst (1995), bu requires
unbounded aspectual situations, and Lin (2003a) maintains that bu selects states. According to
Soh and Gao (2006), the negator bu scopes higher than verbal -le. Instead, negative perfective
sentences are formed with the negator mei and the auxiliary you (50c, Ernst, 1995).
Frequency adverbials cross-linguistically favour imperfective over perfective aspect.
French past tense sentences that contain adverbials like souvent ‘often’ typically use the
Imparfait (51a):
(51)
a.
Saint Louis visitait
souvent les pauvres.
Saint Louis visited.imp often
[French]
the poor
‘Saint Louis often visited the poor.’
b. Dans les semaines qui suivirent on parla souvent et on écrivit beaucoup autour
38
de ce nouveau compagnon.
‘In the weeks that followed.ps we spoke.ps often en wrote.ps a lot around this
novel companion.’
However, examples in which frequency adverbials combine with the Passé Simple are
also found, as illustrated in (51b). They are typically interpreted in relation to a bounded
period of time, as suggested here by dans les semaines qui suivirent. Here the duration
adverbial delimits the frequency situation. Lenci and Bertinetto (2000) offer similar examples
from Italian.
Frequency expressions in Russian generally require the imperfective, as illustrated in
(52a) (from Klimek-Jankowska, 2010). However, within the family of Slavic languages, there
is cross-linguistic variation, for Czech allows perfective aspect in these contexts, as observed
by Eckert (1985, p. 179). Klimek-Kankowska (2010) illustrates this with example (52b):
(52)
a.
Každyi raz
Every
kogda on padal,
on vstaval.
[Russian]
time when he fell.imp, he stood.imp up.
‘Every time he fell, he stood up.’
b. Každé
always
když
spadl,
tak
when perf.fell then
vstal.
[Czech]
perf.stood up
‘Whenever he fell, he stood up.’
According to Fortuin (2008), the choice of imperfective aspect in combination with
frequency adverbials reflects the unbounded nature of the sentence containing a non-specific
number of situations. Klimek-Jankowska (2010) takes the choice of imperfective aspect in
(52a) to mark habituality, whereas the perfective aspect in (52b) reflects the sequential
39
relation between the events. These analyses suggest that the frequency adverbial takes narrow
scope with respect to imperfective aspect in (52a), but wide scope with respect to perfective
aspect in (52b). In languages in which aspect scopes high, only the first option is available (cf.
49 and the discussion in de Swart, 2010).
Further interesting issues arise when we combine aspect-sensitive operators which
make conflicting demands on grammatical aspect. We have already seen in examples (49b)
and (51b) that the French Passé Simple and Imparfait are only sensitive to the highest
aspectual operator in the structure. Given the amalgamation of tense and grammatical aspect
in this language, that outcome is unsurprising. But what about recursive applications of
aspectual operators in Russian and Mandarin? Although measurement phrases may be
compatible with perfective aspect in Russian sentences like (44b), this possibility disappears
when the adverbial measures the duration of a habitual or frequentative situation, so the
imperfective aspect is required in (53).
(53)
Poslednie 10 let
Last
ja
10 years I
kuril
ot
smoke.IMP from
20 do 25 sigaret
v den’. [Russian]
20 to 25 cigarettes in day.
‘For ten years, I smoked between 20 and 25 cigarettes a day. ’
As for Mandarin, bu is not only incompatible with the perfective marker le, but also
with a duration adverbial, as in (54a). As Ernst (1995) points out, the measurement phrase
imposes boundaries upon the process of sleeping, but bu requires an unbounded situation, so
the infelicity of (54a) indicates that bu takes scope over the time adverbial:
(54)
a.
*Jinrong bu shui
Jinrong neg sleep
ba-ge
xiaoshi
eight-cl hour
40
[Mandarin]
b. Jinrong yiban
bu
shui ba-ge
xiaoshi
Jinrong in-general neg sleep eight-cl hour
c.
Ta bu-shi qu le Beijing
he neg-be go asp Beijing
‘It is not the case that he went to Beijing.’
Interestingly, the sentence is fine if we add an adverbial phrase indicating frequency or
habituality (54b). Here bu takes scope over the adverbial yiban, and indicates that Jinrong was
not in the habit of sleeping eight hours a day. These data suggest a very high position for
negation in the aspectual structure of Mandarin. In contrast, Soh and Gao (2006) argue that
verbal -le takes scope below bu-shi ‘not-be’ in examples like (54c).
The examples treated in this section indicate that the layered aspectual structure
introduced in section 4 above opens up new ways of investigating the interaction of
grammatical aspect and other aspect sensitive expressions such as measurement phrases,
negation and frequency adverbials in a wide range of languages.
6. Conclusion
The domain of verbal aspect is complex, because various factors come into play: lexical
features of the verb, the semantics of predicate-argument structure, aspectual operators like
the Progressive and perfective/imperfective aspect, and aspect-sensitive expressions such as
measurement adverbials, negation and frequency adverbs. A layered structure in which tense
scopes over grammatical aspect, which in turn scopes over aspectual class, allows us to study
the interactions between these expressions in a cross-linguistic perspective. The basic
distinction between aspectual class and grammatical aspect is not always easy to establish,
because perfectivity and telicity may interact, as we saw in Russian and Mandarin Chinese.
Similarly, tense and aspect are amalgamated in French, which correlates with differences in
41
the way the perfective/imperfective contrast works out in Slavic and Romance languages.
Clearly, a full-fledged theory of aspect requires a cross-linguistic perspective.
42
References
Arsenijvič, B. (2006). Inner aspect and telicity: The decompositional nature of eventualities
at the syntax-semantics interface. PhD thesis, Leiden University (LOT dissertation
series 142).
Bennett, M., & Partee, B. (1972). Towards the logic of tense and aspect in English. Santa
Monica, CA: System Development Corporation. Reprinted in B. H. Partee, (2004),
Compositionality in formal semantics: Selected papers by Barbara H. Partee. Oxford:
Blackwell.
Bertinetto, P., & Delfitto, D. (2000). Aspect vs. actionality: Why they should be kept apart.
In: Ö. Dahl (Ed.), Tense and aspect in the languages of Europe (pp. 189–226). Berlin:
Mouton de Gruyter.
Binnick, R. I. (1991). Time and the verb: A guide to tense and aspect. New York: Oxford
University Press.
Bohnemeyer, J. & M. Swift (2004). Event realization and default aspect, Linguistics and
Philosophy, 27, 263–296.
Bonami, O. (2002). A syntax-semantics interface for tense and aspect in French. In F. van
Eynde, et al. (Eds.), Proceedings of the HPSG Conference 2001 (pp. 31–50). Stanford:
CSLI Publications.
Bonomi, A. (1997). Aspect, quantification and when-clauses in Italian. Linguistics and
Philosophy, 20, 469–514.
Borik, O. (2002). Aspect and reference time. PhD thesis, Utrecht University (LOT dissertation
series, 64).
Brès, J. (2005). L’imparfait dit narratif. Paris: CNRS-Editions.
Brunot, F., & Bruneau, C. (1949). Précis de grammaire historique de la langue française.
Paris: Masson.
43
Caudal, P. 2005. Stage structure and stage salience for event semantics. In P. Kempchinsky &
R. Slabakova (Eds.), Aspectual inquiries (pp. 239–264). Dordrecht: Springer.
Cipria, A., & Roberts, C. (2000). Spanish imperfecto and preterito: Truth conditions and
Aktionsart effects in a situation semantics. Natural Language Semantics, 8, 297–347.
Comrie, B. (1976). Aspect. Cambridge: Cambridge University Press.
Comrie, B. (1985). Tense. Cambridge: Cambridge University Press.
Dahl, Ö. (1985). Tense and aspect systems. Oxford: Blackwell.
Dahl, Ö. & Velupillai, V. (2008). Perfective/imperfective aspect. In M. Haspelmath, M. S.
Dryer, D. Gil, & B. Comrie (Eds.), The world atlas of language structures online
(chapter
65).
Munich:
Max
Planck
Digital
Library.
Available
online
at
http://wals.info/feature/65, Accessed on February 24, 2010.
Depraetere, I. (1995). On the necessity of distinguishing between (un)boundedness and
(a)telicity. Linguistics and Philosophy, 18, 1–19.
Di Sciullo, A. M., & Slabakova, R. (2005). Quantification and aspect. In H. Verkuyl, H. de
Swart, & A. van Hout (Eds.), Perspectives on aspect (pp. 61–80). Dordrecht: Springer.
Dowty, D. R. (1979). Word meaning and Montague Grammar: The semantics of verbs and
times in Generative Semantics and Montague’s PTQ. Dordrecht: Reidel.
Eckert, E. (1985). Aspect in repetitive contexts in Russian and Czech. In M. S. Flier and A.
Timberlake (Eds.), The scope of Slavic aspect (pp. 169–180). Columbus: Slavica
Publishers.
Ernst, T. (1995). Negation in Mandarin Chinese. Natural Language and Linguistic Theory,
13, 665–707.
Filip, H. (1999). Aspect, situation types and noun phrase semantics. New York: Garland.
44
Filip, H. (2000). The quantization puzzle. In J. Pustejosvky & C. Tenny (Eds.), Events as
grammaticalized objects: From the combined perspective of lexical semantics, logical
semantics and syntax (pp. 3–60). Stanford: CSLI Publications.
Filip, H. (To appear). Aspectual class and Aktionsart. In C. Maienborn, K. von Heusinger &
P. Portner (Eds.), Semantics: An international handbook of natural language meaning.
Berlin: De Gruyter.
Fortuin, E. (2008). Frequency, iteration and quantity: the semantics of expressions of frequent
repetition in Russian and their relationship to aspect. Russian Linguistics, 32, 203–243.
Garey, H. B. (1957). Verbal aspects in French. Language, 33, 91–110.
Gehrke, B. (2007). Ps in motion. PhD thesis, Utrecht University (LOT dissertation series,
184).
Grønn, A. (2003). The semantics and pragmatics of the Russian factual imperfective. PhD
thesis, Oslo University. Published in the series Acta Humaniora, Oslo, 2004.
Jackendoff, R. (1996). The proper treatment of measuring out, telicity and perhaps even
quantification in English. Natural Language and Linguistic Theory, 14, 305–354.
Jonge, B. de (2000). Eventuality classification: Meaning and use of Spanish simple past
tenses. In: E. Contini-Morava & Y. Tobin (Eds.), Between grammar and lexicon (pp.
227–254). Amsterdam, John Benjamins.
Kamp, H. & Reyle, U. (1993). From discourse to logic: Introduction to modeltheoretic
semantics, formal logic and discourse representation theory. Dordrecht: Kluwer.
Kiparsky, P. (1998). Partitive case and aspect. In M. Butt & W. Geuder (Eds.), The projection
of arguments: Lexical and compositional factors (pp. 275–307). Stanford: CSLI
Publications.
45
Klimek-Jankowska, D. (2010). Quantified eventualities in Russian, Czech and Polish: An OT
analysis. In D. Shu & K. Turner (Eds.), Contrasting meaning in the languages of the
East and the West (pp. 283–317). Bern: Peter Lang.
Koenig, J.-P., & Muansuwan, N. (2005). The syntax of aspect in Thai. Natural Language and
Linguistic Theory, 23, 335–380.
Krifka, M. (1989). Nominal reference, temporal constitution and quantification in event
semantics. In: R. Bartsch, J. van Benthem, & P. van Emde Boas (Eds.), Semantics and
contextual expression (pp. 75–115). Dordrecht: Foris.
Krifka, M. (1992). Thematic relations as links between nominal reference and temporal
constitution. In: I. Sag & A. Szabolcsi (Eds.), Lexical matters (pp. 29–53). Stanford,
CA: CSLI Publications.
Landman, F. (1992). The progressive. Natural Language Semantics, 1, 1–32.
Lenci, A., & Bertinetto, P. M. (2000). Aspect, adverbs and events: habituality vs. perfectivity.
In: J. Higginbotham, F. Pianesi, & A. C. Varzi (Eds.), Speaking of events (pp. 245–287).
Oxford: Oxford University Press.
Lin, J.-W. (2003a). Aspectual selection and negation in Mandarin Chinese. Linguistics, 41,
425–459.
Lin, J.-W. (2003b). Temporal reference in Mandarin Chinese. Journal of East Asian
Linguistics, 12, 259–311.
Moens, M., & Steedman, M. (1988). Temporal ontology and temporal reference.
Computational Linguistics, 14(2), 15–28.
Molendijk, A. (1990). Le passé simple et l’imparfait: Une approache reichenbachienne.
Amsterdam: Rodopi.
Parsons, T. (1990). Events in the semantics of English: A study in subatomic semantics.
Cambridge, MA: The MIT Press.
46
Portner, P. (1998). The Progressive in modal semantics. Language, 74(4):760–87.
Portner, P. (2003). The temporal semantics and the modal pragmatics of the perfect. Linguistics
and Philosophy, 26, 459–510.
Pustejovsky, J. (1995). The generative lexicon. Cambridge, MA: The MIT Press.
Ramchand, G. (2008). Verb meaning and the lexicon: A first phase syntax. Cambridge:
Cambridge University Press.
Reichenbach, H. (1947). Elements of symbolic logic. New York: Free Press. Paperback
edition, 1966.
Ritter, E., & Rosen, S. T. (2005). Topic or aspect: Functional heads, features and the
grammatization of events. In P. Kempchinsky & R. Slabokova (Eds.), Aspectual
inquiries (21-40). Dordrecht: Kluwer.
Ritz, E. (In press). The perfect crime? Illicit uses of the present perfect in Australian police
media releases. Journal of Pragmatics.
Schaden, G. (2009). Present perfects compete. Linguistics and Philosophy, 32, 115–141.
Schoorlemmer, M. (1995). Participial passive and aspect in Russian. PhD thesis, Utrecht
University.
Schmitt, C. (2001). Cross-linguistic variation and the present perfect: The case of Portuguese.
Natural Language and Linguistic Theory, 19, 403–453.
Smith, C. (1991/1997). The parameter of aspect. Dordrecht: Kluwer.
Smith, C., & Erbaugh, M. S. (2005). Temporal interpretation in Mandarin Chinese.
Linguistics, 43, 713–756.
Soh, H. L. (2009) Speaker presupposition and Mandarin Chinese sentence-final -le: a unified
analysis of the “change of state” and the “contrary to expectation reading. Natural
Language and Linguistic Theory, 27, 623–657.
47
Soh, H. L., & Gao, M. (2006). Perfective aspect and transition in Mandarin Chinese: An
analysis of double -le sentences. In P. Denis, E. McCready, A. Palmer, & B. Reese
(Eds.), Proceedings of 2004 Texas Linguistics Society Conference: Issues at the
semantics-pragmatics interface (pp. 107–122). Somerville, MA: Cascadilla Press.
Soh, H. L., & Kuo, J. Y.-C. (2005). Perfective aspect and accomplishment situations in
Mandarin Chinese. In A. van Hout, H. de Swart, and H. Verkuyl (Eds.), Perspectives on
aspect (pp. 199–216). Dordrecht: Springer.
Swart, H. de. (1998). Aspect shift and coercion. Natural Language and Linguistic Theory, 16,
347–385.
Swart, H. de. (2007). A cross-linguistic discourse analysis of the perfect, Journal of
Pragmatics, 39, 2273–2307.
Swart, H. de. (2010). Circonstantiels temporels et aspect verbal: interactions dans les
contextes épisodiques et habituels. In: N. Flaux, D. Stosic, & J. Vet (Eds.), Interpréter
les temps verbaux (pp. 83–105). Bern: Peter Lang.
Swart, H. de. (To appear). Mismatches and coercion. In C. Maienborn, K. von Heusinger , &
P. Portner (Eds.). Semantics: an international handbook of natural language meaning.
Berlin: De Gruyter.
Swart, H. de, & Molendijk, A. (1999). Negation and the temporal structure of narrative
discourse. Journal of Semantics, 16, 1–42.
Sybesma, R. (1999). The Mandarin VP. Dordrecht: Kluwer.
Tasmowski-De Ryck, L. (1985). L’imparfait avec et sans rupture. Langue française, 67, 59–
77.
Travis, L. (2000). Event structure in syntax. In: C. Tenny & J. Pustejovsky (Eds.), Events as
grammatical objects: The converging perspectives of lexical semantics and syntax (pp.
145–185). Stanford, CA: CSLI Publications.
48
Travis, L. (2010). Inner aspect: The articulation of VP. Dordrecht: Springer.
Van Geenhoven, V. (2005). Atelicity, pluractionality and adverbial quantification. In H.
Verkuyl, H. de Swart, & A. van Hout (Eds.), Perspectives on aspect (pp. 107–124).
Dordrecht: Springer.
Vendler, Z. (1957). Verbs and times. Philosophical Review, 66(2), 143–160. Reprinted in Z.
Vendler, (1967), Linguistics in philosophy (pp. 97–121). Ithaca: Cornell University
Press.
Verkuyl, H. J. (1972). On the compositional nature of the aspects. Dordrecht: Reidel.
Verkuyl, H. J. (1993). A theory of aspectuality: The interaction between temporal and
atemporal structure. Cambridge: Cambridge University Press.
Verkuyl, H. J. (1999). Aspectual issues. Stanford, CA: CSLI Publications.
Vlach, F. (1981). The semantics of the progressive. In P. Tedeschi & A. Zaenen (Eds.), Tense
and aspect (Syntax and Semantics 14) (pp. 271–292). New Yord: Academic Press.
Xiao, R., & McEnery, T. (2004). Aspect in Mandarin Chinese: A corpus-based study.
Amsterdam: John Benjamins.
Zucchi, S. (1999). Incomplete events, intensionality and imperfective aspect. Natural
Language Semantics, 7, 179–215.
49
1
The Editor would like to thank Jadranka Gvozdanovic for correcting some typos in a
preliminary draft of this chapter.
50